Your opinion (UNIX,VMS)
Jakob Erber
don't be afraid, this is not another "what is better discussion", but I
would be interested what you think about this point:
Most big companies especially in the finance area, but also all kind of
other businesses still run there core business on big Host (IBM, UNISYS...)
systems and do not indent to change this. If in rare cases they do, then VMS
comes into consideration as replacement, not UNIX. Is it really correct to
compare UNIX and VMS like competitors? In many cases they certainly compete,
but I would say: VMS on Alpha is the only true alternative of Mini (Middle?)
Computers to Mainframes, concerning availablility. UNIX has its place in
other areas.
I know a bit, what I'm talking about, because my company successfully
replaced a big UNISYS Mainframe of an important bank by a large VMS Cluster.
What do you think?
Jakob
JF Mezei
> Most big companies especially in the finance area, but also all kind of
> other businesses still run there core business on big Host (IBM, UNISYS...)
> systems and do not indent to change this. If in rare cases they do, then VMS
> comes into consideration as replacement, not UNIX.
Sorry. But the big banks are all tying UNIX based applications to their
mainframes. For instance, they don't run the telephone-banking on their
mainframes, they runb it on some Unix box that sends transactiosn to their
mainframes. And their web servers run on Sun solaris, agains with connector
software to send transactions to the mainframes. And there is plenty of NT
servers as well in banks doing lots of stuff.
If you are fortunate enough to work for a bank that embraced VMS, then good
for you. You should consider youself lucky.
Christof Brass
Good news, thanks!
For replacing mainframes and minis I fully agree that there
aren't many alternatives and from the perspective of power and
stability VMS comes in at the first place.
But there are other areas where UNIX and VMS compete whereas the
VMS position is weakened by the fact that there are not all
necessary apps available.
VMS can be used as web hosting platform like UNIX. This is an
area where it can compete with UNIX.
Hamlyn Mootoo
(today) would be that there is an application that only runs on VMS
which is better suited to running their business than any they can find
elsewhere, OR that the IT manager/CIO/CFO is comfortable with VMS having
used it elsewhere. To make a decision based on pure hardware/software
platform, to Alpha/VMS, otherwise would be a somewhat insane decison by
any savvy businessperson given the fact that future support of both
hardware and software is at best, dubious. What some of my colleagues
fail to realize here is that although we all share an extreme fondness
for VMS, from a business standpoint, it does not make especially good
sense. Computers are used by businesses to get work done more
efficiently than by doing it by hand. A computer system shares the same
rank as a copier in that it is a TOOL. If you were going to buy a
copier, would you go with a current one, or one that the manufacturer
has all but stated is at end of life? There are many mid-range computer
systems based on UNIX which are quite stable and highly available from
vendors such as IBM and HP, who offer such products as HA (highly
available) clusters for AIX, and MC Service Guard for HP UX (among other
tools), in a price range very competitive with Alpha/VMS clustering. It
is not our lack of faith or enthiusiasm as Sys Admins. which cause VMS
to wither, it is the lack of marketing and support by certain
individuals in the very company who invented it , that is to blame. DEC
had some of the best hardware and software engineers in the world
working for it, and quite possibly the world's worst sales and marketing
people in history. If DEC gets (posthumously) into the Guiness World
Records book, it will sadly be for how the
ignorant-of-the-product-line-and-technology marketing and sales people
sank a company in record time.
HM
mulp
"Hamlyn Mootoo" <uni...@bigfoot.com> wrote in message
news:3B35369C...@bigfoot.com...
> to wither, it is the lack of marketing and support by certain
> individuals in the very company who invented it , that is to blame. DEC
> had some of the best hardware and software engineers in the world
> working for it, and quite possibly the world's worst sales and marketing
> people in history.
I think you miss the real problem that still handicaps VMS:
the DEC unix strategy did not embrace VMS
Sun developed an entend, embrace, engulf strategy. Sun funded NFS on VMS
A/D work which became the basis for every VMS TCP stack except for UCX.
This allowed Sun to go into a VMS shop and say "VMS is great, VAXclusters
are great, we love VMS, but our workstations are cheaper and they work fine
with your VMS servers." When someone said, we can't switch to unix because
we have this code written in FORTRAN using VMS system services, Sun
responded, "No problem, here's Acceler8 to get you up and running".
On the other hand, the DEC unix mindset was, "the best thing is for DEC is
to kill off VMS, but if the company won't do it, then we'll just tell our
customers that they are idiots if they keep doing things on VMS". Hey, it
was Lucente, the worst VP that DEC ever had, that told Oracle that they
should stop developing on VMS and switch to unix.
VMS was/is tied to the well being of the architecture that it runs on.
While the PDP-8 and PDP-11 familes had strong support from "partners"
because they not only could add value by integrating solutions for
customers, they were absolutely required because of all the things that DEC
didn't supply. In fact, DECUS was founded and supported by DEC for decades
because KO, et al, knew that DEC would be successful by making its customers
dependent on DEC being successful.
Then in the mid-80s, the attitude changed and lots of DEC's "partners" got
screwed as DEC cut them out of the value chain.
Today the question is whether Microsoft is a good partner. Hey, this is
just another page in the DEC playbook that Microsoft has adapted for their
play. One key change that Microsoft made was to embrace their partners, and
then when they became too profitable, engulf them. Usually, Microsoft
engulfed them by buying a cut rate license for an old version of their
software and then bundled it; ultimately it kills off the partner's product.
but at least the partner got a pot of money to start up a new venture or to
retire. If DEC had bought the TCP stack from Kasten, he might not have as
motivated to form TGV to offer a competing product; even better, a smart
deal would have been to encourage him to deliver a lot of missing pieces to
complement the product. I argued for years that UCX should not include any
support for mail; later when I learned of PMDF, I argued that a license for
"PMDF-lite" should be negotiated and bundled with UCX so that all complaints
about Internet mail could be answered with "talk to Innosoft: they can sell
you a solution".
Rdb is an interesting case. There were several database products on VMS
(dating to the PDP-11), but they weren't exploiting VMS fully. If the
reason for doing Rdb was to exploit VMS, then it might have been justified,
but the real reason was to grab money from Ingres and Oracle. Needless to
say, they weren't too happy and their sales reps would redirect customers to
other, friendlier computer vendors. DEC gave Rdb to Oracle for free, if not
less, but the result has been to tied a significant portion of Oracle's
profit to the DEC product line. Oracle is a master of holding two
contradictory thoughts: IBM is the enemy because of DB2; IBM is our friend
because IBM's enemy, Microsoft, is Oracle' enemy. VMS and Oracle should get
alone fine since they share common friends and common enemies: Microsoft and
Microsoft, for example. Windows is clearly a fact of life that needs to be
embraced, but still, Microsoft it the enemy.
Years ago the minicomputer was declared dead, but IBM took a mishmash of
minis that were more confusing than all the operating systems for the PDP-11
and -8 combined, consolidated them, and then targeted them at VAX/VMS.
AS/OS/400 the VAX killer, then became the unix killer. No, it doesn't
generally go into unix shops and display unix, but it does go into DEC, IBM,
Unisys, HP shops and replace what might have become unix systems. The enemy
is Sun and for every Sun solution, IBM has a set of partners who can deliver
an IBM solution; in fact, many of the IBM partners are also Sun partners.
Of course, after alienating all of its partners, forcing them to focus on
Windows or unix, DEC then started selling off its software. Lots of
critical system management software went to CA which did what CA does,
increase prices until there's no more money to be made and then force the
customers who can't go elsewhere to move to Windows or Solaris. This
pattern was repeated so that DEC was no longer a solution where a year
earlier it was. Note this didn't affect just VMS, but also Alpha unix and
wnt. Of course, alienating DEC unix customers was SOP with all the jerking
around that ISVs refused to deal with, killing VAX support and then killing
MIPS support being just two of the bad moves.
Still, IBM has demonstrated that you can buck "common wisdom", the common
wisdom being that IBM is a dinosaur with no future. And VMS has likewise
failed to follow the common wisdom, and the resulting marketing plan, by
failing to decline in revenue and profit contribution so that it could be
discarded without notice.
Christof Brass
This is true. I worked two years for a company having a first
class storage management solution (for real goods not HW) and
beeing at the same time DEC OEM for certain components that DEC
didn't offer, e.g. a controller for disk mirroring. HP tried to
persuade the company to port their SW to HP/UX and succeeded
only when DEC started to talk against VMS.
> Today the question is whether Microsoft is a good partner. Hey, this is
> just another page in the DEC playbook that Microsoft has adapted for their
> play. One key change that Microsoft made was to embrace their partners, and
> then when they became too profitable, engulf them. Usually, Microsoft
> engulfed them by buying a cut rate license for an old version of their
> software and then bundled it; ultimately it kills off the partner's product.
> but at least the partner got a pot of money to start up a new venture or to
> retire. If DEC had bought the TCP stack from Kasten, he might not have as
> motivated to form TGV to offer a competing product; even better, a smart
> deal would have been to encourage him to deliver a lot of missing pieces to
> complement the product. I argued for years that UCX should not include any
> support for mail; later when I learned of PMDF, I argued that a license for
> "PMDF-lite" should be negotiated and bundled with UCX so that all complaints
> about Internet mail could be answered with "talk to Innosoft: they can sell
> you a solution".
Micro$hit has two other strategies to kill a major competitor:
a) the soft one: buy the company like the one which produced
Visio. b) the hard one: develop a product that can compete and
sell it for much less or bundle it with the OS, e.g. Word
(against WordPerfect) or IE (against Navigator). Because
Micro$hit has that much money they could afford selling
something at a dumping price for a long time, longer than the
competitor can survive without revenue from cut in market share.
> Rdb is an interesting case. There were several database products on VMS
> (dating to the PDP-11), but they weren't exploiting VMS fully. If the
> reason for doing Rdb was to exploit VMS, then it might have been justified,
> but the real reason was to grab money from Ingres and Oracle. Needless to
> say, they weren't too happy and their sales reps would redirect customers to
> other, friendlier computer vendors. DEC gave Rdb to Oracle for free, if not
> less, but the result has been to tied a significant portion of Oracle's
> profit to the DEC product line. Oracle is a master of holding two
> contradictory thoughts: IBM is the enemy because of DB2; IBM is our friend
> because IBM's enemy, Microsoft, is Oracle' enemy. VMS and Oracle should get
> alone fine since they share common friends and common enemies: Microsoft and
> Microsoft, for example. Windows is clearly a fact of life that needs to be
> embraced, but still, Microsoft it the enemy.
Offering more than the plain OS is a serious problem for every
vendor. A similar problem arose when Digital tried to transform
itself into an independent consulting company. Most people
didn't believe that Digital's consultant would propose a
non-Digital solution.
> Years ago the minicomputer was declared dead, but IBM took a mishmash of
> minis that were more confusing than all the operating systems for the PDP-11
> and -8 combined, consolidated them, and then targeted them at VAX/VMS.
> AS/OS/400 the VAX killer, then became the unix killer. No, it doesn't
> generally go into unix shops and display unix, but it does go into DEC, IBM,
> Unisys, HP shops and replace what might have become unix systems. The enemy
> is Sun and for every Sun solution, IBM has a set of partners who can deliver
> an IBM solution; in fact, many of the IBM partners are also Sun partners.
Honestly, I don't see how the AS/OS/400 was a VAX killer. This
system is extreme expensive, slow and almost impossible to
program efficiently. And as far as I know there is even no X11
available.
> Of course, after alienating all of its partners, forcing them to focus on
> Windows or unix, DEC then started selling off its software. Lots of
> critical system management software went to CA which did what CA does,
> increase prices until there's no more money to be made and then force the
> customers who can't go elsewhere to move to Windows or Solaris. This
> pattern was repeated so that DEC was no longer a solution where a year
> earlier it was. Note this didn't affect just VMS, but also Alpha unix and
> wnt. Of course, alienating DEC unix customers was SOP with all the jerking
> around that ISVs refused to deal with, killing VAX support and then killing
> MIPS support being just two of the bad moves.
>
> Still, IBM has demonstrated that you can buck "common wisdom", the common
> wisdom being that IBM is a dinosaur with no future. And VMS has likewise
> failed to follow the common wisdom, and the resulting marketing plan, by
> failing to decline in revenue and profit contribution so that it could be
> discarded without notice.
IBM have a future but in HW.